WELCOME TO METABOX 2.0

A toolbox for thorough metabolomic data analysis, integration and interpretation





Example data sets

Data processing and analysis pipeline

GCGC_DM - plasma samples from diabetes mellitus cohort: 15 Pooled-QC samples and 2 ISs, 60 samples x 89 metabolites, see [Ref]

GC_Milk - retail milk samples: 10 Pooled-QC samples and 1 IS, 87 samples x 16 fatty acids, see [Ref]

LC_LN - urine samples from lupus nephritis cohort: 1 IS, 116 samples x 8 metabolites, see [Ref]

Data integration pipeline

GC_Fat_Tissue and LC_Fat_Tissue - adipose tissue samples from colorectal carcinoma cohort, metabolites are measured by 2 platforms, see [Ref]

Data interpretation pipeline

metabolite_list - table of metabolites, p-values, and fold-changes for ORA or enrichment analysis

metabolite_protein_list - list of metabolites and proteins for integrated pathway analysis

About

Metabox 2.0

Metabox 2.0: A toolbox for thorough metabolomic data analysis, integration and interpretation. Metabox 2.0 is an updated version of the R package Metabox , released in 2016. The tool includes several methods for data processing, statistical analysis, biomarker analysis, integrative analysis and data interpretation. This GUI supports a wide range of users, from bench biologists to experienced bioinformaticians. It comes with an intuitive web interface for simple data analysis. We recommend the R command line version for custom pipelines and other exclusive projects.

Updates

version 2.6 (JUNE 2023)
  • Change report location
  • Set default package color
  • Update MUVR to current version
  • Fix bug when running univariate analysis
version 2.5 (MAY 2023)
  • Update pathway data for enrichment analysis
version 2.4 (MAR 2023)
  • Add example data sets for GUI version
version 2.3 (FEB 2023)
  • Summarize coefficient of variation (cv) and normality
  • Fix default scaling of PCA plot
  • Fix default color
version 2.2 (OCT 2022)
  • Fix bug when running MUVR
version 2.1 (SEP 2022)
  • Add imputation methods: zero, half-min
  • Allow scaling and block weighting by the block inertia
  • Allow missing values in normalization, transformation and scaling
version 2.0 (JUL 2022)
  • Initial release
References

Metabox 2.0 - A toolbox for thorough metabolomic data analysis, integration and interpretation, see [Ref]

Metabox 1.0 - A toolbox for metabolomic data analysis, interpretation and integrative exploration was released in 2016, see [Ref]




PCA score plot of the 1st two PCs provides the overview of the data.

                                        

                                      

                                    

                                    

                                      

                                      

                                      

                                        

                                        

                                      

                                      

                                        
*Drag on the plot to select sample(s), double-click to unselect.

                                        
*To exclude sample(s)/outlier(s) in further analysis, click REMOVE SAMPLE.

                                        

                                          

                                        

                                        
                                        

                                      

                                    

                                  

                                  

                                    

                                    

                                      
RLA plot provides the overview of the replicates within each group.

                                      
*Display only the first 100 samples.

                                      

                                      

                                    

                                    

                                      

                                    

                                  

                                  

                                    

                                    

                                      
Density plot provides the distribution of variables and samples.

                                      

                                    

                                    

                                      

                                        

                                      

                                      

                                        

                                      

                                    

                                    

                                      

                                      

                                        

                                      

                                      

                                        

                                      

                                    

                                  

                                  

                                    

                                    
                                    

                                    
*To exclude variable(s) in further analysis, select column(s) and click DELETE COLUMN.

                                    

                                    

                                      

                                    

                                    

                                  

                                

                              

                            

                          

                        

                      

                    

                  

                

                

                  

                    

                      

                        

                          
                        

                      

                      

                        

                          

                          

                            
                            

                              

                                
                                

                                  

                                    

                                    

                                      
PCA score plot of the first two PCs provides the overview of the data.

                                      

                                      

                                        

                                      

                                      

                                        

                                          

                                        

                                      

                                    

                                    

                                      

                                      

                                      

                                        

                                        

                                      

                                      

                                        
*Drag on the plot to select sample(s), double-click to unselect.

                                        
*To exclude sample(s)/outlier(s) in further analysis, click REMOVE SAMPLE.

                                        

                                          

                                        

                                        
                                        

                                      

                                    

                                  

                                  

                                    

                                    

                                      
RLA plot provides the overview of the replicates within each group.

                                      
*Display only the first 100 samples.

                                      

                                      

                                    

                                    

                                      

                                    

                                  

                                  

                                    

                                    

                                      
Density plot provides the distribution of variables and samples.

                                      

                                    

                                    

                                      

                                        

                                      

                                      

                                        

                                      

                                    

                                    

                                      

                                      

                                        

                                      

                                      

                                        

                                      

                                    

                                  

                                  

                                    

                                    
                                    

                                    
*To exclude variable(s) in further analysis, select column(s) and click DELETE COLUMN.

                                    

                                    

                                      

                                    

                                    

                                  

                                

                              

                            

                          

                        

                      

                    

                  

                

              

            

          

          

            

              
              

                

                  

                    

                      

                        

                          
                        

                      

                      

                        

                          

                          

                            
                            

                              

                                
                                

                                  

                                    

                                    

                                      
Statistical significance plot provides the -log10(adjusted p-value) of each variable.
                                                                                                                               Dashed line represents statistical significance cutoff (adjusted p-value < 0.05).

                                      
*Display only the top 100 variables, sorte by adjusted p-values. Click on a dot to toggle its boxplot.

                                      

                                      

                                    

                                    

                                      

                                    

                                  

                                  

                                    
*For 2-level factor/class, fold change (fc) = level1/level2, for more than 2 levels, fc = each_level/all_mean. Statistical significance for post hoc test is adjusted p-value < 0.05.

                                    

                                      

                                    

                                    

                                  

                                

                              

                            

                          

                        

                      

                    

                  

                

                

                  

                    

                      

                        

                          
                        

                      

                      

                        

                          

                          

                            
                            

                              

                                
                                

                                  

                                    

                                    

                                      
Score and loading plots of 2 principal components (PC).

                                      

                                      

                                        

                                          
                                          

                                            
                                            
                                          

                                        

                                      

                                      

                                        

                                          
                                          

                                            
                                            
                                          

                                        

                                      

                                      

                                    

                                    

                                      

                                      

                                    

                                    

                                      

                                      

                                    

                                  

                                  

                                    

                                    
Loading plot of a principal component (PC).

                                    
*Display only the top 100 variables, sorted by loading values.

                                    

                                    

                                      
                                      

                                        
                                        
                                      

                                    

                                    

                                      

                                    

                                  

                                  

                                    

                                    

                                      
Variable importance in projection (VIP) plot.

                                      
*Display only the top 100 variables, sorted by VIP values.

                                      

                                      

                                    

                                  

                                  

                                    

                                    
Output table contains loading values of variables in each PC and VIP values for PLS and OPLS.

                                    

                                    

                                      

                                    

                                    

                                  

                                

                              

                            

                          

                        

                      

                    

                  

                

                

                  

                    

                      

                        

                          
                        

                      

                      

                        

                          

                          

                            
                            

                              

                                
                                

                                  

                                    

                                    

                                      
*Drag on the plot to zoom, double-click to reset.

                                      
*Plotting will take awhile for a lot of variables.

                                      

                                      

                                    

                                  

                                  

                                    

                                    
Output table contains correlation coefficient, p-value and adjusted p-value between two variables.

                                    

                                    

                                      

                                    

                                    

                                  

                                

                              

                            

                          

                        

                      

                    

                  

                

                

                  

                    

                      

                        

                          
                        

                      

                      

                        

                          

                          

                            
                            

                              

                                
                                

                                  

                                    

                                    
Output table contains Intercept, coefficients and chi-square test results for significant fixed effects.

                                    

                                    

                                      

                                    

                                    

                                  

                                

                              

                            

                          

                        

                      

                    

                  

                

              

            

          

          

            

              

                

                  

                    
                  

                

                

                  

                    

                    

                      
                      

                        

                          
                          

                            

                              

                              

                                

                              

                            

                            

                              

                              

                                

                              

                            

                            

                              

                              
Output table contains order and average VIP rank of variables in minimal-optimal model.